This chapter covers the distinct roles of Power BI Desktop and Power BI Service in the Microsoft Power BI ecosystem, a critical topic for the DP-900 exam. Understanding the difference between these two components is essential because exam questions frequently test your ability to identify which tool to use for specific tasks—creation vs. sharing, data modeling vs. dashboard consumption. Approximately 10-15% of the exam questions touch on Power BI concepts, with a significant portion focusing on the Desktop vs. Service distinction. By the end of this chapter, you will be able to accurately describe the purpose, capabilities, and limitations of each tool, and choose the correct one for given scenarios.
Jump to a section
Think of building a Power BI report like building a custom car. Power BI Desktop is your fully equipped garage—it has all the tools: an engine lift (Power Query), a diagnostic computer (DAX), a paint booth (visual formatting), and a welding rig (data modeling). You can spend hours in the garage, tweaking the engine, adjusting the suspension, and customizing the interior. No one sees the car until you're ready. But the garage is a single-user environment; only you can work on the car at one time. Now, Power BI Service is the public road system. Once the car is built, you drive it out of the garage and onto the road. On the road, the car can be driven by anyone with a license (a Power BI Pro or Premium license). The road has traffic rules (row-level security), speed limits (dashboard performance), and service stations (scheduled data refresh). The Service also allows multiple drivers to take the wheel simultaneously—colleagues can view dashboards, create reports from the same dataset, or set up alerts. However, you cannot rebuild the engine while driving on the highway; major modifications must go back to the garage. In this analogy, the garage (Desktop) is for creation and heavy lifting, while the road (Service) is for consumption, sharing, and collaboration. A finished report is like a car that is parked in a shared garage (workspace) on the Service, ready for others to drive.
What is Power BI Desktop?
Power BI Desktop is a free, client-side application installed on a Windows PC. It is the primary authoring tool for creating data models, transforming data, and building reports. It is not a cloud service; it runs locally and stores work in .pbix files. The Desktop connects to hundreds of data sources—both on-premises (SQL Server, Excel, CSV) and cloud (Azure SQL Database, Salesforce, Google Analytics). It includes Power Query Editor for data transformation and the DAX formula language for creating calculated columns, measures, and tables. Desktop is single-user: only one person can open and edit a .pbix file at a time. It does not support real-time collaboration or sharing directly; you must publish to the Service to share.
What is Power BI Service?
Power BI Service is a cloud-based Software as a Service (SaaS) offering hosted in Microsoft Azure. It is accessed via a web browser (app.powerbi.com) and provides capabilities for publishing, sharing, and consuming reports and dashboards. The Service supports collaboration through workspaces, where multiple users can co-own and manage content. It enables scheduled data refresh (up to 8 times per day with Pro, 48 with Premium), row-level security (RLS), and natural language queries (Q&A). Dashboards—a Service-only feature—allow pinning visualizations from multiple reports onto a single canvas. The Service also includes apps (packaged content) and allows embedding reports in websites or Teams. Unlike Desktop, the Service is not a data modeling tool; you cannot create measures or transform data natively. All data modeling must be done in Desktop before publishing.
Key Differences in Capabilities
#### Data Modeling and Transformation - Desktop: Full Power Query Editor for merging, appending, pivoting, and cleaning data. DAX for calculated columns, measures, and tables. Relationship creation and management between tables. - Service: No Power Query Editor. No ability to create DAX measures or columns. You can edit existing measures in the model view (if you have Build permission) but only after the model is published. The Service does not support creating new tables or relationships from scratch.
#### Sharing and Collaboration - Desktop: Only one user at a time. Sharing a .pbix file is manual—email, OneDrive, or network share. No version control built-in. - Service: Workspaces allow multiple users to collaborate. You can assign roles: Admin, Member, Contributor, Viewer. Reports and dashboards can be shared with individuals, groups, or the entire organization. Apps distribute curated content to many users.
#### Data Refresh - Desktop: Manual refresh only. You can click "Refresh" in Desktop, but it does not schedule automatic refreshes. - Service: Scheduled refresh via the dataset settings. You must configure a gateway for on-premises data sources. The Service supports up to 8 daily refreshes with Pro license, 48 with Premium per dataset. Additionally, the Service supports DirectQuery and Live Connection modes that query the source in real-time.
#### Security - Desktop: No row-level security (RLS) enforcement. You can define RLS roles in Desktop, but they are only enforced after publishing to the Service. - Service: RLS is enforced when users view reports. You can also configure app workspace permissions and share dashboards with specific users.
#### Performance and Scalability - Desktop: Limited by local machine resources (RAM, CPU). Large datasets (over 1 GB) may cause performance issues. Desktop uses VertiPaq in-memory engine, but the entire dataset must fit in RAM. - Service: Leverages Azure cloud infrastructure. Premium capacity offers dedicated resources, larger dataset limits (up to 100 GB per dataset in Premium Gen2), and higher refresh frequencies. The Service also supports dataflows and datamarts for ETL in the cloud.
Publishing from Desktop to Service
To share a report, you publish the .pbix file from Desktop to the Service. The process: 1. In Desktop, click Publish in the Home ribbon. 2. Select a target workspace in the Service (My Workspace or a shared workspace). 3. The report and its dataset are uploaded. The dataset appears in the workspace's datasets list. 4. After publishing, you can create dashboards, set up refresh, and share.
When to Use Each
Use Desktop for: initial data exploration, data cleaning, modeling, creating complex measures, and building reports from scratch.
Use Service for: sharing reports and dashboards, scheduling data refresh, collaborating in workspaces, applying RLS, creating dashboards from multiple reports, and consuming content on mobile devices.
Exam Traps
Trap: Thinking you can create a dashboard in Desktop. Dashboards are a Service-only feature. Desktop creates reports; reports have pages, dashboards have tiles.
Trap: Believing Desktop supports real-time collaboration. Desktop is single-user; the Service supports co-authoring via workspaces.
Trap: Assuming the Service can transform data. The Service has no Power Query; all transformation must be done in Desktop or via dataflows.
Trap: Confusing "report" and "dashboard" — a report is a multi-page document from Desktop; a dashboard is a single-page canvas in the Service.
Interaction with Related Technologies
Power BI Report Builder: A separate tool for creating paginated reports (print-optimized). Paginated reports are published to the Service and require Premium.
Power BI Mobile Apps: Consume content from the Service on phones and tablets. They cannot create reports.
Power BI Embedded: Allows developers to embed Service reports into custom applications. Requires an Azure capacity.
Excel Integration: You can analyze Power BI datasets in Excel using the Analyze in Excel feature, which connects to the Service.
Default Values and Limits
Pro license: 10 GB per user storage, 1 GB per dataset (uncompressed), 8 daily refreshes.
Premium per user (PPU): 100 TB storage, 100 GB per dataset, 48 daily refreshes.
Premium capacity: varies by SKU (e.g., P1: 100 TB storage, 100 GB per dataset, 48 refreshes).
Desktop: free, no storage limit on local files, but performance degrades with large datasets.
The Service automatically compresses data using VertiPaq. The compression ratio varies; typical is 5-10x.
Step-by-Step Publishing Verification
After publishing, you can verify the content in the Service:
Go to app.powerbi.com.
Navigate to the target workspace.
Check the Reports tab to see the published report.
Check the Datasets tab to see the dataset.
Click on the dataset to configure refresh settings.
Create a dashboard by pinning visuals from the report.
Install Power BI Desktop
Download the latest version from the Microsoft Store or the Microsoft Download Center. The installation is straightforward—run the .exe and follow prompts. After installation, open the application. You'll see a blank canvas with three main areas: the ribbon at top, the canvas in the middle, and the Visualizations/Fields panes on the right. No license is required for Desktop; it's free. However, some premium features (e.g., deployment pipelines) require a Pro or Premium license to use when publishing.
Connect to a Data Source
Click 'Get Data' in the Home ribbon. Choose from hundreds of connectors: SQL Server, Excel, CSV, Azure Blob Storage, etc. For a SQL Server database, you'll need the server name and database name. You can choose Import (default) or DirectQuery. Import copies data into the model; DirectQuery queries the source live. For Import, you can apply transformations in Power Query Editor. After selecting tables, click 'Load' to bring data into the model. The data is compressed using VertiPaq and stored in memory.
Build a Report in Desktop
Drag fields from the Fields pane onto the canvas. Choose a visual type from the Visualizations pane (bar chart, table, map, etc.). Create measures using DAX: for example, `Total Sales = SUM(Sales[Amount])`. Build relationships between tables by dragging fields in the Model view. Add slicers, filters, and drillthrough pages. You can also add bookmarks and buttons for interactivity. Save the file as a .pbix file.
Publish to Power BI Service
Click 'Publish' in the Home ribbon. Select a target workspace—either 'My Workspace' (personal) or a shared workspace. If you don't have a workspace, you can create one in the Service beforehand. Desktop uploads the .pbix file, which includes the report and the dataset. The dataset is extracted and stored in the Service. After publishing, a success dialog appears with a link to open the report in the Service.
Configure Scheduled Refresh
In the Service, navigate to the dataset settings. Under 'Scheduled refresh', toggle to enable. If the data source is on-premises, you must install and configure an On-premises Data Gateway. Select the gateway cluster. Set the refresh frequency (e.g., daily at 8 AM and 5 PM). For cloud sources (e.g., Azure SQL), no gateway is needed. The Service will refresh the dataset according to the schedule, pulling new data from the source. With Pro, you can schedule up to 8 refreshes per day; with Premium, up to 48.
Share and Collaborate in Service
Once published, you can share the report with others. Click 'Share' on the report and enter user email addresses. You can also create a dashboard by pinning visuals from the report. For collaboration, create a workspace and add members with appropriate roles (Admin, Member, Contributor, Viewer). Members can edit reports and datasets. You can also create an app to distribute a collection of reports and dashboards to a large audience. Row-level security (RLS) is enforced in the Service based on roles defined in Desktop.
Enterprise Scenario 1: Sales Reporting for a Global Company
A multinational corporation with 10,000 employees uses Power BI to provide sales dashboards to regional managers. The data team uses Power BI Desktop to connect to an on-premises SQL Server database containing sales transactions. They use Power Query to clean and transform the data, create measures like Year-over-Year Growth, and build a comprehensive report with drillthrough pages for each region. The .pbix file is published to a dedicated workspace in Power BI Service. A gateway is installed on a server in the corporate network to enable scheduled refresh every 4 hours. Regional managers access the report via the Service or the mobile app. Row-level security is configured so that managers only see data for their region. The report is also embedded in SharePoint Online for easy access. The key challenge is managing the gateway cluster for high availability—if the gateway goes down, refreshes fail, and the data becomes stale. The team also uses deployment pipelines (a Premium feature) to move content from dev to test to production.
Enterprise Scenario 2: Self-Service Analytics with Excel Integration
A mid-sized company encourages self-service analytics. Business analysts use Power BI Desktop to create their own reports from a shared dataset published to the Service. The dataset is created by the central IT team using Power BI Desktop and published to a workspace. Analysts can connect to this dataset using the 'Live Connection' option in Desktop, which means they build reports without duplicating data. They can also use 'Analyze in Excel' to pivot the dataset in Excel. The Service handles the data refresh, so analysts always work with current data. The challenge is governance: IT must manage dataset permissions and ensure that analysts don't accidentally publish conflicting reports. They use the Service's lineage view to track dependencies and certifications to mark trusted datasets.
Scenario 3: Real-Time Dashboard for Operations
A logistics company needs a real-time dashboard showing package tracking. They use Power BI Desktop with a streaming dataset (e.g., from Azure Stream Analytics). In Desktop, they create a report with real-time visuals (e.g., line chart updating every second). They publish the report to the Service, which supports streaming datasets natively. The dashboard is displayed on a large screen in the operations center. The Service allows them to set alerts (e.g., if package volume exceeds a threshold). The challenge is that streaming datasets have limited storage (1 million rows) and cannot be combined with imported data in the same dataset. They must use a separate dataset for historical analysis.
What DP-900 Tests on Power BI Desktop vs. Service
The DP-900 exam objective 3.4 is "Describe Power BI capabilities and integration with Azure data services." Within this, the exam specifically tests the ability to distinguish between the tools used for creation (Desktop) and sharing (Service). You will not be asked to write DAX or configure a gateway, but you must know:
Which tool is used for data transformation (Desktop).
Which tool supports dashboards (Service only).
Which tool supports scheduled refresh (Service).
Which tool supports real-time collaboration (Service via workspaces).
The licensing requirements: Desktop is free; Service requires Pro or Premium for sharing and scheduled refresh.
Common Wrong Answers and Why Candidates Choose Them
"You can create dashboards in Power BI Desktop." This is false. Dashboards are a Service-only feature. Candidates confuse dashboards with report pages. In Desktop, you create reports with multiple pages. In the Service, you pin visuals from one or more reports to a dashboard. The exam may show a screenshot of a dashboard and ask which tool was used to create it—the answer is the Service, not Desktop.
"Power BI Service can transform data using Power Query." False. The Service has no Power Query Editor. Candidates think that because the Service can edit datasets (e.g., change measure formulas), it can also transform data. But transformation (splitting columns, merging queries) is only in Desktop or via Dataflows (a separate Service feature). The exam may ask: "Which tool would you use to merge two queries?" Answer: Desktop.
"Power BI Desktop can be used to share reports with colleagues." False. Desktop is a single-user tool. You can share the .pbix file manually, but that is not a direct sharing feature. The exam expects you to know that sharing is done via the Service. A question might say: "You need to share a report with 50 users. What should you do?" Correct answer: Publish to the Service and share from there.
"Scheduled data refresh is configured in Power BI Desktop." False. Scheduled refresh is a Service feature. Desktop only supports manual refresh. Candidates might see the 'Refresh' button in Desktop and assume it can be scheduled. The exam may ask: "How often can you schedule a refresh in Power BI Pro?" Answer: Up to 8 times per day.
Specific Numbers and Terms That Appear on the Exam
Pro license: 8 daily scheduled refreshes, 10 GB per user storage, 1 GB per dataset limit.
Premium: 48 daily refreshes, 100 GB per dataset (Gen2), dedicated capacity.
Desktop: free, no storage limit, but performance depends on local hardware.
Gateway: required for on-premises data sources in the Service.
Workspace roles: Admin, Member, Contributor, Viewer.
Dashboard: single page, can have tiles from multiple reports.
Report: multi-page, created in Desktop.
Edge Cases the Exam Loves to Test
What if you need to edit a measure after publishing? You can edit it in the Service if you have Build permission, but the original .pbix in Desktop is out of sync. The exam may ask: "You modified a measure in the Service. What happens to the original .pbix file?" Answer: It is not updated; you must republish to overwrite.
Can you use the Service without a license? Yes, you can sign up for a free trial, but to view shared content you need a Pro or Premium license (unless the content is in a Premium capacity and shared with free users). The exam may ask: "A user with a free license can view reports in a Premium workspace." True.
Can you create a dashboard from a published report? Yes, by pinning visuals. But you cannot create a dashboard from Desktop.
How to Eliminate Wrong Answers
Always ask: Is this a creation task or a consumption task? If the task involves building, modeling, or transforming, the answer is Desktop. If it involves sharing, scheduling, or consuming, the answer is Service. If the question mentions 'dashboard', it's Service. If it mentions 'report page', it could be either, but dashboards are Service-only.
Power BI Desktop is for authoring reports and data modeling; Power BI Service is for sharing, collaboration, and consumption.
Dashboards are exclusive to Power BI Service; reports are created in Desktop and published to the Service.
Scheduled data refresh is configured in the Service, not in Desktop. Desktop supports only manual refresh.
Data transformation (Power Query) is only available in Desktop or Dataflows (Service), not directly in the Service.
Power BI Desktop is free with no license required; the Service requires a Pro or Premium license for sharing and scheduled refresh.
Row-level security (RLS) roles are defined in Desktop but enforced only after publishing to the Service.
Pro license allows up to 8 scheduled refreshes per day; Premium allows up to 48.
A gateway is required for on-premises data sources when using scheduled refresh in the Service.
These come up on the exam all the time. Here's how to tell them apart.
Power BI Desktop
Free client application installed locally on Windows.
Primary tool for data transformation using Power Query Editor.
Supports DAX for creating measures, calculated columns, and tables.
Single-user environment; no direct sharing or collaboration.
Manual refresh only; no scheduled refresh capability.
Power BI Service
Cloud-based SaaS accessed via web browser (app.powerbi.com).
No Power Query Editor; data transformation not possible natively.
Can edit existing measures but cannot create new ones from scratch.
Multi-user collaboration via workspaces with role-based access.
Supports scheduled refresh (up to 8/day Pro, 48/day Premium) and gateways.
Mistake
Power BI Desktop can create dashboards.
Correct
Power BI Desktop creates reports with multiple pages. Dashboards are a feature of Power BI Service only. A dashboard is a single-page canvas that can contain tiles pinned from one or more reports. There is no dashboard creation capability in Desktop.
Mistake
Power BI Service can transform data using Power Query.
Correct
Power BI Service does not include Power Query Editor. Data transformation must be done in Power BI Desktop before publishing, or via Power BI Dataflows (a separate service) that uses Power Query Online. The Service can only edit existing measures and relationships in the model view.
Mistake
You need a Pro license to use Power BI Desktop.
Correct
Power BI Desktop is completely free to download and use. No license is required to create reports, connect to data, or build models. However, to publish to the Service and share content, you need a Power BI Pro or Premium license. Some advanced features (e.g., deployment pipelines) also require a license.
Mistake
Scheduled data refresh is configured in Power BI Desktop.
Correct
Scheduled refresh is configured in the Power BI Service on the dataset settings page. Desktop only supports manual refresh (clicking Refresh). The Service uses a gateway for on-premises sources and can refresh up to 8 times per day with Pro, 48 with Premium.
Mistake
Power BI Service allows multiple users to edit the same report simultaneously.
Correct
The Service supports co-authoring in workspaces, but only one user can edit a report at a time. If two users open the same report in edit mode, the second user sees a read-only notice. True real-time co-authoring (like Google Docs) is not supported. Users can collaborate by working in separate reports or using the service's commenting feature.
Reveal each answer, then mark whether you got it right. Score 60%+ to unlock the next chapter.
Power BI Desktop is a free, local Windows application for creating data models and reports. Power BI Service is a cloud-based platform for sharing, collaborating, and consuming reports and dashboards. Desktop is for authoring; Service is for distribution. You cannot create dashboards or schedule data refreshes in Desktop; those are Service features.
No. Dashboards are a feature exclusive to Power BI Service. Desktop creates reports, which are multi-page documents. In the Service, you can pin visuals from one or more reports to a single-page dashboard. The term 'dashboard' in Desktop context is a common exam trap.
No. Power BI Desktop is free to download and use. However, to publish reports to the Service and share them, you need a Power BI Pro or Premium license. Some features like deployment pipelines require Premium. You can still create and save .pbix files locally without any license.
Scheduled refresh is configured in Power BI Service. After publishing a dataset, go to the dataset settings, find 'Scheduled refresh', toggle it on, and set the frequency. For on-premises data sources, you must install and configure an On-premises Data Gateway. Desktop does not support scheduled refresh.
Not simultaneously. In the Service, if one user opens a report in edit mode, others see a read-only notice. True co-authoring is not supported. Users can collaborate by working in separate reports or using the commenting feature. Desktop is strictly single-user.
Power BI Pro allows up to 8 scheduled refreshes per day per dataset. Premium capacities allow up to 48 per day. This is a frequently tested number on the DP-900 exam.
No. Power Query Editor is only available in Power BI Desktop. The Service does not have a native data transformation interface. However, you can use Power BI Dataflows (a separate feature) which provides Power Query Online for data preparation in the cloud.
You've just covered Power BI Service vs Power BI Desktop — now see how well it sticks with free DP-900 practice questions. Full explanations included, no account needed.
Done with this chapter?